From here rsyslog puts it in the different log files, e.g. Home Assistant running in Docker makes use of the journald log driver, which means that all the Home Assistant STDOUT and STDERR output ends up in the system journal. Until the feature request is implemented there is an alternative to persist logs. Makes reading the log a bit easier to follow especially if there are built in Delays or Waits. Script initiations are prefixed as Script -Īnd service actions are either Set - or Get. Triggered events in automations are prefixed as Action. T08:42:01.035220+00:00 Get - T2 set point was at 79.įor clarity I use a simple nomenclature scheme
#Docker syslog rotation code#
Its incredibly easy to set up, completely controlled by you and can be toggled on/off with a simple input_boolean (assuming you place the notify_xxx AFTER the relevant code block.įinally, here is a small bit of the data logged today: Home Assistant notifications (Log started: T14:34:51.008347+00:00)
#Docker syslog rotation update#
I get logging to home-assistant.log as per everyone else plus way more to /var/log/syslog (with the default logger: - nothing is set in configuration.yaml): :~$ grep hass /var/log/syslog|tail -10Īpr 12 10:57:38 JonLabNUC hass: 10:57:38 INFO (MainThread) Updating component: sensor.motion_sensor_158d000221a1ad_batteryĪpr 12 10:57:38 JonLabNUC hass: 10:57:38 INFO (MainThread) battery_sensor_from_attributes: Executing step call serviceĪpr 12 10:57:40 JonLabNUC hass: 10:57:40 INFO (MainThread) Executing battery_sensor_from_attributesĪpr 12 10:57:40 JonLabNUC hass: 10:57:40 INFO (MainThread) battery_sensor_from_attributes: Running scriptĪpr 12 10:57:40 JonLabNUC hass: 10:57:40 INFO (MainThread) battery_sensor_from_attributes: Executing step call serviceĪpr 12 10:57:40 JonLabNUC hass: 10:57:40 INFO (MainThread) Component has already been discovered: sensor plug_158d0002431039_battery, sending updateĪpr 12 10:57:40 JonLabNUC hass: 10:57:40 INFO (MainThread) Got update for entity with hash: ('sensor', 'plug_158d0002431039_battery') '. I’m running under venv on ubuntu 16.04 LTS, as a service Path to PID file useful for running as daemonĮnables daily log rotation and keeps up to the v, -verbose Enable verbose logging to file.
#Docker syslog rotation install#
skip-pip Skips pip install of required packages on startup open-ui Open the webinterface in a browser debug Start Home Assistant in debug mode safe-mode Start Home Assistant in safe mode c path_to_config_dir, -config path_to_config_dirĭirectory that contains the Home Assistant
version show program's version number and exit h, -help show this help message and exit Home Assistant: Observe, Control, Automate. I’m starting hass in a venv with: hass -log-rotate-days 4 I am not a serious coder, but willing to help and test. Now many people, including me, often say in response to requests like this “get coding” or “where is your PR?”. Proper logging over a network via syslog. This can probably be done easily by a user via command_line. Include pastebinit by default in home assistant core images, and a service to use pastebinit to post your logs to a paste site and return a url to it. Optionally the older files could be zipped to save space. On restart of home assistant the homeassistant.log is renamed to homeassistant.log.1, homeassistant.log.1 is renamed to homeassistant.log.2, and so on up to the maximum set in your configuration.
I propose that there be a user configurable number of logs to keep. So valuable information that might help debugging is lost. The problem I perceive is that the log disappears on each start of home assistant. No criticism, I suspect the current system probably arose out of small SD card and not wanting to over fill the thing. Coming from the perspective of a long time linux user, I find logging in home assistant a bit in need of improvement.
0 kommentar(er)
